Artificial Intelligence

Artificial
Intelligence I:
introduction
Lecturer: Tom Lenaerts
Institut de Recherches Interdisciplinaires et de
Développements en Intelligence Artificielle (IRIDIA)
Université Libre de Bruxelles

TLo (IRIDIA) 2October 13, 2015
Practical stuff
 Course homepage:
 http://iridia.ulb.ac.be/~tlenaert/teach/ai1.
html
 Mailinglist:
 02452-valvas@elvas.vub.ac.be
 Textbook:
 S. Russell and P. Norvig Artificial
Intelligence:A Modern Approach
Prentice Hall, 2003, Second Edition
 Excercices:
 Joachim de Beule en Bart De
Vylder.
 Lisp (~scheme++)
 Exam: written at end of 1st. sem.
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.

Course overview
 What is AI.
 Intelligent agents.
 Problem solving.
 Knowledge and reasoning.
 Planning.
 Uncertain knowledge and reasoning.
 Learning.
 Communicating, perceiving and acting.

Outline
 What is AI
 A brief history
 The State of the art (see book)
 Reading
 Lisp
 And its relation to Scheme

What is Artificial Intelligence
 Creative extension of philosophy:
 Understand and BUILD intelligent entities
 Origin after WWII
 Highly interdisciplinary
 Currently consist of huge variety of subfields
 This course will discuss some of them

What is Artificial Intelligence
 Different definitions due to different criteria
 Two dimensions:
 Thought processes/reasoning vs. behavior/action
 Success according to human standards vs. success
according to an ideal concept of intelligence:
rationality.
Systems that think like humans Systems that think rationally
Systems that act like humans Systems that act rationally

Systems that act like humans
 When does a system behave intelligently?
 Turing (1950) Computing Machinery and Intelligence
 Operational test of intelligence: imitation game
 Test still relevant now, yet might be the wrong question.
 Requires the collaboration of major components of AI:
knowledge, reasoning, language understanding, learning,
…

Systems that act like humans
QuickTime™ and a
QuickTime™ and a
Andrew Hodges.
Alan Turing, the enigma
Available at amazon.co.uk
Problem with Turing test: not reproducible, constructive or
amenable to mathematical analysis.

Systems that think like humans
 How do humans think?
 Requires scientific theories of internal brain activities
(cognitive model):
 Level of abstraction? (knowledge or circuitry?)
 Validation?
 Predicting and testing human behavior
 Identification from neurological data
 Cognitive Science vs. Cognitive neuroscience.
 Both approaches are now distinct from AI
 Share that the available theories do not explain anything
resembling human intelligence.
 Three fields share a principal direction.

Systems that think like humans
 Some references;
 Daniel C. Dennet.
Consciousness explained.
 M. Posner (edt.)
Foundations of cognitive
science
 Francisco J. Varela et al.
The Embodied Mind
 J.-P. Dupuy. The
mechanization of the mind
QuickTime™ and a
QuickTime™ and a

Systems that think rationally
 Capturing the laws of thought
 Aristotle: What are ‘correct’ argument and thought
processes?
 Correctness depends on irrefutability of reasoning processes.
 This study initiated the field of logic.
 The logicist tradition in AI hopes to create intelligent systems using
logic programming.
 Problems:
 Not all intelligence is mediated by logic behavior
 What is the purpose of thinking? What thought should one have?

Systems that think rationally
 A reference;
 Ivan Bratko, Prolog
programming for
artificial intelligence.
QuickTime™ and a

Systems that act rationally
 Rational behavior: “doing the right thing”
 The “Right thing” is that what is expected to maximize
goal achievement given the available information.
 Can include thinking, yet in service of rational action.
 Action without thinking: e.g. reflexes.
 Two advantages over previous approaches:
 More general than law of thoughts approach
 More amenable to scientific development.
 Yet rationality is only applicable in ideal environments.
 Moreover rationality is not a very good model of reality.

Systems that act rationally
 Some references;
 Michael Wooldridge.
Reasoning about
rational agents.
QuickTime™ and a

Rational agents
 An agent is an entity that perceives and acts
 This course is about designing rational agents
 An agent is a function from percept histories to actions:
 For any given class of environments and task we seek the
agent (or class of agents) with the best performance.
 Problem: computational limitations make perfect
rationality unachievable.
f:P*→A

Foundations of AI
 Different fields have contributed to AI in the form of
ideas,viewpoints and techniques.
 Philosophy: Logic, reasoning, mind as a physical system,
foundations of learning, language and rationality.
 Mathematics: Formal representation and proof algorithms,
computation, (un)decidability, (in)tractability, probability.
 Psychology: adaptation, phenomena of perception and motor
control.
 Economics: formal theory of rational decisions, game theory.
 Linguistics: knowledge represetatio, grammar.
 Neuroscience: physical substrate for mental activities.
 Control theory: homeostatic systems, stability, optimal agent
design.

A brief history
 What happened after WWII?
 1943: Warren Mc Culloch and Walter Pitts: a model of
artificial boolean neurons to perform computations.
 First steps toward connectionist computation and learning
(Hebbian learning).
 Marvin Minsky and Dann Edmonds (1951) constructed the first
neural network computer
 1950: Alan Turing’s “Computing Machinery and
Intelligence”
 First complete vision of AI.

A brief history (2)
 The birth of AI (1956)
 Darmouth Workshop bringing together top minds on automata
theory, neural nets and the study of intelligence.
 Allen Newell and Herbert Simon: The logic theorist (first nonnumerical
thinking program used for theorem proving)
 For the next 20 years the field was dominated by these participants.
 Great expectations (1952-1969)
 Newell and Simon introduced the General Problem Solver.
 Imitation of human problem-solving
 Arthur Samuel (1952-)investigated game playing (checkers ) with great
success.
 John McCarthy(1958-) :
 Inventor of Lisp (second-oldest high-level language)
 Logic oriented, Advice Taker (separation between knowledge and
reasoning)

A brief history (3)
 The birth of AI (1956)
 Great expectations continued ..
 Marvin Minsky (1958 -)
 Introduction of microworlds that appear to require intelligence to
solve: e.g. blocks-world.
 Anti-logic orientation, society of the mind.
 Collapse in AI research (1966 - 1973)
 Progress was slower than expected.
 Unrealistic predictions.
 Some systems lacked scalability.
 Combinatorial explosion in search.
 Fundamental limitations on techniques and representations.
 Minsky and Papert (1969) Perceptrons.

A brief history (4)
 AI revival through knowledge-based systems (1969-
1970)
 General-purpose vs. domain specific
 E.g. the DENDRAL project (Buchanan et al. 1969)
 First successful knowledge intensive system.
 Expert systems
 MYCIN to diagnose blood infections (Feigenbaum et al.)
 Introduction of uncertainty in reasoning.
 Increase in knowledge representation research.
 Logic, frames, semantic nets, …

A brief history (5)
 AI becomes an industry (1980 - present)
 R1 at DEC (McDermott, 1982)
 Fifth generation project in Japan (1981)
 American response …
 Puts an end to the AI winter.
 Connectionist revival (1986 - present)
 Parallel distributed processing (RumelHart and
McClelland, 1986); backprop.

A brief history (6)
 AI becomes a science (1987 - present)
 Neats vs. scruffies.
 In speech recognition: hidden markov models
 In neural networks
 In uncertain reasoning and expert systems: Bayesian network
formalism
 …
 The emergence of intelligent agents (1995 - present)
 The whole agent problem:
“How does an agent act/behave embedded in real environments with
continuous sensory inputs”

Reading
 One course on AI provides a
limited view on a vast research
area
 Reading assignment:
 P.1-138
 Provides background
information on the current
perspective in AI: embodied
cognitive science.

Lisp vs. scheme
 Lisp (= LISt Processor) is the second
oldest programming language still in use
(after FORTRAN).
 Invented by John McCarthy at MIT in
1958.
 Until the mid '80s Lisp was more a family
of dialects than a single language.
 In 1986 an ANSI subcommittee was
formed to standardize these dialects into a
single Common Lisp.
 The result being the first Object Oriented language to
become standardized, in 1994.
 Once you understand Common Lisp it is
easy to adapt yourself to weaker dialects as
for instance Scheme.

Lisp vs Scheme
 Lisp has much more built-in functions and special forms, the Scheme
language definition takes 45 pages while Common Lisp takes 1029
pages)
 Apart from lexical variables(lexically scoped) Lisp also has special
variables (dynamically scoped)
 In a lexically scoped language, the scope of an identifier is fixed
at compile time to some region in the source code containing
the identifier's declaration. This means that an identifier is only
accessible within that region (including procedures declared
within it).
 In a dynamically scoped language an identifier can be referred
to, not only in the block where it is declared, but also in any
function or procedure called from within that block, even if the
called procedure is declared outside the block

Lisp vs Scheme
 Statically vs dynamically scoped variables
>(set 'regular 5)
5
>(defun check-regular ()
regular)
CHECK-REGULAR
>(check-regular)
5
> (let ((regular 6))
(check-regular))
5
>(defvar *special* 5)
*SPECIAL*
>(defun check-special ()
*special*)
CHECK-SPECIAL
>(check-special)
5
>(let ((*special* 6))
(check-special))
6

Lisp vs Scheme
 Scheme evaluates the function part of a function call in exactly the same
way as arguments, Lisp doesn’t.
 In Lisp, the role of the symbol depends on the position in the list
(fun arg)
 Example:
(let ((list '(1 2 3)))
(list list))
==>((1 2 3))
 Function calls: Scheme vs. Lisp
(let ((fun (compute-a-function)))
(fun x y)
(map car L)
let ((fun (compute-a-function)))
(funcall fun x y))
(map ‘list #’car L)

Lisp vs Scheme
 Scheme uses one name space for functions, variables, etc.,
Lisp doesn’t.
 Special (global) symbols
(defun square (x) …)
(setf (symbol-function square) (x) …)
 Lexical (local) symbols
(labels ((square (x) …)) …)
(setf square (function (lambda (x) …))

Lisp vs Scheme
 In Lisp defun defines functions in the global
environment even if the function is defined internally to
another function.
(define (stack)
(let ((data ‘()))
(define (push elm) …)
(define (pop) …)
…)
(defun stack ()
(let ((data ‘()))
(defun push (elm) …)
(defun pop () …)
…)
(defun stack ()
(let ((data ‘()))
(labels ((push (elm) …)
(pop () …))
…)

Lisp vs Scheme
 Lisp functions can have rest and optional parameters.
Scheme functions only can have the equivalent of a rest
parameter.
((lambda (a b) (+ a (* b 3))) 4 5) => 19
((lambda (a &optional (b 2)) (+ a (* b 3))) 4 5) => 19
((lambda (a &optional (b 2)) (+ a (* b 3))) 4) => 10
((lambda (&optional (a 2 b) (c 3 d) &rest x) (list a b c d x)))
=> (2 nil 3 nil nil)
((lambda (&optional (a 2 b) (c 3 d) &rest x) (list a b c d x)) 6)
=> (6 t 3 nil nil)

Lisp vs Scheme
 Lisp functions can have also keyword parameters.
 Or some mixture.
((lambda (a b &key c d) (list a b c d)) 1 2) => (1 2 nil nil)
((lambda (a b &key c d) (list a b c d)) 1 2 :c 6) => (1 2 6 nil)
((lambda (a b &key c d) (list a b c d)) 1 2 :d 8) => (1 2 nil 8)
((lambda (a b &key c d) (list a b c d)) 1 2 :c 6 :d 8) => (1 2 6 8)
((lambda (a b &key c d) (list a b c d)) 1 2 :d 8 :c 6) => (1 2 6 8)
((lambda (a &optional (b 3) &rest x &key c (d a))
(list a b c d x)) 1) => (1 3 nil 1 ())
(list a b c d x)) 1 2) => (1 2 nil 1 ())
(list a b c d x)) :c 7) => (:c 7 nil :c ())

Lisp vs Scheme
 Lisp has standard macros, Scheme since R5RS.
 “Lisp macros are a way to execute arbitrary code at "compile
time", using entities that are called like functions, but evaluate
their arguments (or not, if they choose not to) in ways that are
controlled by the macro itself. The language used to write the
macro is just Lisp itself, so the full power of the language is
available”
 Allows you to define your own special forms as if or
and.

Lisp vs Scheme
 Lisp has special forms (loop, do, dotimes, …) for
looping, in Scheme the user is asked to use tail-
recursion that is implemented efficiently.
(loop for i fixnum from 3
when (oddp i) collect i
while (< i 5))
⇒ (3 5)
http://iridia.ulb.ac.be/~tlenaert/prog/Lisp/cltl/cltl2.html

Other courses at the VUB
 AI does not end here …
 Artificiele Intelligentie II
 Technieken van de AI I en II
 Autonomous Agents
 Adaptive Systems I en II
 Machine Learning
 Multi-agent systems
 …

Some references
 Understanding Intelligence by Rolf
Pfeifer and Christian Scheier.
 Artificial Intelligence: Structures
and Strategies for Complex
Problem-solving by George Luger.
 Computation and Intelligence:
Collective readings edited by
George Luger.
 Paradigms of Artificial Intelligence
Programming: Case Studies in
Common Lisp by Peter Norvig.

Artificial Intelligence

More Related Content

What's hot

Viewers also liked

Similar to Artificial Intelligence

More from Mhd Sb

Recently uploaded

Artificial Intelligence